Why not try to predict autism spectrum disorder with crucial biomarkers in cuproptosis signaling pathway?

Abstract

The exact pathogenesis of autism spectrum disorder (ASD) is still unclear, yet some potential mechanisms may not have been evaluated before. Cuproptosis is a novel form of regulated cell death reported this year, and no study has reported the relationship between ASD and cuproptosis. This study aimed to identify ASD in suspected patients early using machine learning models based on biomarkers of the cuproptosis pathway. We collected gene expression profiles from brain samples from ASD model mice and blood samples from humans with ASD, selected crucial genes in the cuproptosis signaling pathway, and then analysed these genes with different machine learning models. The accuracy, sensitivity, specificity, and areas under the receiver operating characteristic curves of the machine learning models were estimated in the training, internal validation, and external validation cohorts. Differences between models were determined with Bonferroni's test. The results of screening with the Boruta algorithm showed that FDX1, DLAT, LIAS, and ATP7B were crucial genes in the cuproptosis signaling pathway for ASD. All selected genes and corresponding proteins were also expressed in the human brain. The k-nearest neighbor, support vector machine and random forest models could identify approximately 72% of patients with ASD. The artificial neural network (ANN) model was the most suitable for the present data because the accuracy, sensitivity, and specificity were 0.90, 1.00, and 0.80, respectively, in the external validation cohort. Thus, we first report the prediction of ASD in suspected patients with machine learning methods based on crucial biomarkers in the cuproptosis signaling pathway, and these findings may contribute to investigations of the potential pathogenesis and early identification of ASD.

Keywords: artificial neural network; autism spectrum disorder; biomarkers; cuproptosis; machine learning.

FIGURE 1

The flowchart of the present study. ASD, autism spectrum disorder; KNN, k-nearest neighbor; NB, naive Bayesian; SVM, support vector machine; RF, random forest; ANN, artificial neural network.

FIGURE 2

The visualization of gene expression profiles. The gene expression data were normalized between arrays of mouse brain (A) and human blood (B). The crucial genes in the cuproptosis signaling pathway and the selected arrays of each dataset were visualized in a heatmap (C). ASD, autism spectrum disorder.

FIGURE 3

The expression of selected genes in the brain. Based on the Human Protein Atlas, FDX1, DLAT, LIAS, and ATP7B could all be detected in 12 brain regions. In the A-431 cell line, FDX1 protein and DLAT protein were located in mitochondria (A,B), and LIAS protein was located in mitochondria and the nucleoplasm (C). ATP7B protein was also expressed in the Golgi apparatus in the CACO-2 cell line (D). The schematic graph shows the main location of each protein in cells (E). The target proteins, nuclei and microtubules were stained green, blue, and red, respectively. nTPM, normalized transcript expression values.

FIGURE 4

The performance of each machine learning model. The relation between the number of neighbors (k value) and accuracy in KNN (A). The degree and coefficient of SVM are shown in panel (B). In RF, the interrelation between the number of trees and model error is shown in panel (C). After 43,703 steps, the error was 0.009257 in the ANN model (D). KNN, k-nearest neighbor; SVM, support vector machine; RF, random forest; ANN, artificial neural network.

FIGURE 5

The ROC analysis of each model in the internal validation and external validation cohorts. The X-axis and Y-axis represent specificity and sensitivity, respectively. The AUC values are indicated in the blue area, including KNN model (A,B), Naive Bayes model (C,D), SVM model (E,F), RF model (G,H), and ANN model (I,J). The value of the cut-off point is shown at the inflection point. ROC, receiver operating characteristic; AUC, area under the curve; KNN, k-nearest neighbor; SVM, support vector machine; RF, random forest; ANN, artificial neural network.

FIGURE 6

The evaluation of each model after cross validation. The accuracy and Kappa value are shown in panel (A). Bonferroni’s test results are shown in panel (B). The median numbers are represented by dots, and lines indicate the confidence level. ROC, receiver operating characteristic; AUC, area under the curve; KNN, k-nearest neighbor; SVM, support vector machine; RF, random forest; ANN, artificial neural network.